Bond-Order Approaches to Chemical Complexity: Liquid Carbon and Diamond Surface Reconstructions


Steve Stuart



Potentials based on the bond-order approach have enjoyed considerable success in modeling covalent materials, including hydrocarbons and carbon-based materials.  This talk will highlight a few cases where empirical bond-order potentials have been used effectively, and in which direct comparison can be made with first-principles methods.  Results will be presented for liquid carbon, in which the existence of a liquid-liquid phase transition has been proposed.  No evidence for the liquid-liquid phase transition is found, in agreement with shorter quantum mechanical simulations.  The improvement of the treatment of torsional interactions and empirical modeling of conjugation in the current generation of bond-order potentials allows the accurate modeling of this complex system.  Another success is the modeling of surface reconstructions on diamond surfaces.  Structures of recently proposed graphitic overlayers are examined, in order to assess their possible existence at partial surface hydrogenation.  The simulations indicate interesting properties for these novel structures, and provide better agreement with experimental observations than previous DFT studies. Finally, some speculations will be provided for the future direction of empirical bond-order potentials, and the way that they can be recombined with recent results for analytical bond-order potentials.